Summary:
This is a radical change to the approach KWin used to have. So far KWin
fall back to XRender compositing if it detected llvmpipe. With this
change KWin will use OpenGL compositing also on llvmpipe.
This has drawbacks: it's CPU emulated which we actually do not want.
But the question is whether it still makes sense to keep the no-GL
enforcement.
There are several things to consider here:
* QtQuick will use llvmpipe anyway (including things like Plasmashell)
* glamor might be used in the xorg layer which also uses llvmpipe on
XRender
* more consistent look
* we get OpenGL working in tests (important for e.g. openQA)
* hardware which does not support proper GL gets less, there are
GL drivers for KVM, there are GL drivers for rPI, etc. We don't
need to block it anymore because of possibly bad results on those
hardware. Also they mostly use glamor.
* there is still the way to manually select XRender/QPainter compositing
Reviewers: #kwin, #plasma
Subscribers: kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D2184
-use qstringliteral only when necessary (i.e. not in concat or comparison)
-use qbytearray instead of qstring when dealing with latin1 input and output (glplatform)
-use qstringref to extract numbers from strings (glplatform)
-define qt_use_qstringbuilder to optimize all string concatenations
-anidata: use ctor init lists, add windowType member initialization
REVIEW: 125933
On wayland OpenGL might get initialized before we have a connection
to the XServer and we might only have an xcb connection. So let's
guard the access.
Remove the manually written GL dispatch code, and use libepoxy
to resolve functions.
The only exceptions are GLX_MESA_swap_control, which is not in
the XML API registry, and GL_ARB_robustness/GL_EXT_robustness.
For the latter we want to resolve the functions to the same names
on both GLES and desktop GL, and plug in our own implementations
when the extension is not supported.
The used platform interface is passed to the ::detect method as an
argument. The value is now cached in the GLPlatform instance so that
there is an easy way to figure out which platform interface is currently
used.
Adds a cleanup() method to GLPLatform which deletes the static instance.
It gets invoked from cleanupGL() in kwinglutils on tear-down of the
OpenGL backend. Thus the backend doesn't leave incorrect GL info around.
E.g. if KWin starts with llvmpipe the recommended compositor is XRender.
So the backend and the platform is created. Now KWin starts the XRender
Compositor and the GLPlatform is still valid and when trying to load the
Blur or Contrast effect this will succeed as the GLPlatform data would
allow it. But on first try to use the Effect it will fail due to no
valid OpenGL context and even crash.
REVIEW: 117480
It's no longer needed as the OpenGLBackend has a direct rendering
detection and the remaining OpenGL2 backend aborts if the backend
uses indirect rendering. Thus the GLPlatform must always have a
direct rendering context.
REVIEW: 116829
KWin already has a de facto OpenGL 2 dependency through QML. Combined
with the fact that the OpenGL 1 backend is basically unmaintained and
also unused, it's better to remove it for the new major release.
This change includes:
* Removal of cmake option KWIN_BUILD_OPENGL_1_COMPOSITING
* Removal of KWIN_HAVE_OPENGL_1 compile option and all code
ifdef'ed with it (partially removal of if-else constructs)
* Removal of CompositingType::OpenGL1Compositing (flags are kept
as a core flag should get introduced)
* Driver recommendation for OpenGL1Compositing changed to XRender
(should be evaluated whether the drivers can provide GL2)
* Removal of configuration option "GLLegacy"
* Removal of fooMatrix function in kwinglutils
* Removal of ARBBlurShader
* Removal of legacy code path in GLVertexBuffer
* Removal of GLShaderManager::disable
* if-blocks with ShaderManager::instance()->isValid() removed
REVIEW: 116042
This is a workaround fo QTBUG-34898 and affects the VirtualBox driver
and SandyBridge Mobile. OpenGL compositing is just not possible and
crashes as soon as there is anything rendered with QtQuick. This change
should be reverted once the Qt bug is fixed.
To nevertheless use OpenGL one can as always use the KWIN_COMPOSE env
variable, though this will result in crashes. An alternative is to set
QT_OPENGL_NO_SANITY_CHECK which forces Qt into using the threaded
rendering. At least for Sandybridge this seems to be a workable solution
as it's only causing flickering in fullscreen and KWin doesn't use any
fullscreen QtQuick elements.
* "" needs to be wrapped in QStringLiteral
* QString::fromUtf8 needed for const char* and QByteArray
* QByteArray::constData() needed to get to the const char*
This method returns true when glMapBufferRange() is likely to perform
worse than glBufferSubData() when updating an unused range in a buffer
object.
This is the case with the NVIDIA driver, where glMapBufferRange()
will force thread serialization. The driver tracks which ranges of
the buffer are in use, so calls to glBufferSubData() should not
cause a pipeline stall.
Use glGetStringi() to list the extensions when the GL version is 3.0
or greater. glGetString() does not accept the GL_EXTENSIONS token
in an OpenGL core context.
Add an option to kcmcompositing in the 'Advanced' tab, to enable or
disable color correction. It is specified that it's experimental and it
needs Kolor Manager.
Before painting for a particular screen, ColorCorrection::setupForOutput
should be called.
A screen property is added for WindowPaintData.
In kwinglutils, The fragment shaders are intercepted before being
compiled and they get a couple of lines of code inserted in order to do
the color correction. This happens only when color correction is enabled, of
course.
For D-Bus communication with KolorServer, everything is async.
The implementation basically manages a set of color lookup tables for
different outputs and for different window regions. These are taken via
D-Bus. Each lookup table has around 700 KB.
This commit reintroduces the changes from the former merge with the
"color2" branch. In this form, it can be easily reverted.
REVIEW: 106141
This merge is incomplete and it does not include the review number of
the associated review request. It should have been pushed as a single
commit, because the merged commits were not intended to be published in
their form.
This reverts commit dcba90263069a221a5489b1915c5cf1ca39d090c, reversing
changes made to 50ae07525c7fde07794e7548c3d6e5a69cb1a89d.
Conflicts:
kwin/scene_opengl.cpp
kwin/scene_opengl.h
Adjustment to reality. KWin has had a dependency on Mesa 8.0 for quite
some time given that it is what basically all distributions ship.
It is better to clearly state what is required. For KWin Mesa 8.0 is a
more reliable dependency as all DRI drivers which do not support DRI2
have been removed.
Packagers have been informed about this intended change some weeks ago.
REVIEW: 106799
During detecting the OpenGL capabilities also a recommended compositor
can be set. This recommendation is honoured by the OpenGL based
compositors. That is the SceneOpenGL2 requires a recommendation for at
least OpenGL2 and the SceneOpenGL1 requires at least a recommendation for
OpenGL1. If the driver recommends XRender compositing the SceneOpenGL
performs the existing fallback to XRender.
With this recommendation the hacks in the Scene are removed, e.g. it is
no longer checked whether the driver is software emulation as that is
provided through the recommendation.
To overrule the recommendation the environment variable KWIN_COMPOSE is
extended by the values O1 and O2 to enforce either OpenGL 1 or OpenGL 2.
This overwrites all other checks. As a side-effect this allows now to run
KWin on the llvmpipe:
LIBGL_ALWAYS_SOFTWARE=1 KWIN_COMPOSE=O2 kwin --replace &
But not that I would recommend to use it :-)
REVIEW: 106741
VMware uses a Gallium driver which means that OpenGL based compositing
works out of the box without any adjustments to GLPlatform.
Just adding recognizing code for the SVGA3D driver.
REVIEW: 106826
OpenGL is properly working if there is a direct rendering context.
If LIBGL_ALWAYS_INDIRECT is set VirtualBox falls back to Mesa's software
rasterizer. So in order to get OpenGL the driver is now whitelisted in
the opengltest.
GLPlatform is extended to recognize the VirtualBox driver and has new
methods to report whether it is a virtual machine and VirtualBox. The
detection is rather limited as we don't get access to the underlying
hardware, so we do not know whether the features are really supported.
We need to trust the driver here in announcing the right extensions.
The driver does not provide glxQueryDrawable although it is part of
GLX 1.3. A hack is added in the glxbackend to set the function pointer to
NULL. This can unfortunately not be done in glxResolveFunctions() as
QueryDrawable seems not to be provided by an extension (at least not
listed in the OpenGL registry) and getProcAddress resolves a function but
it only prints an OpenGL Warning to stderr.
As a note: the driver reports that it is using XSHM for
GLX_EXT_texture_from_pixmap.
REVIEW: 106821
The Egl backend is decoupled from the OpenGL ES build option which makes
it possible to use it as a replacement for glx.
To make this possible a new build flag is added when egl is available at
compile time and any egl specific code is now ifdefed with this flag
instead of the gles flag. In addition at runtime a windowing system enum
value is passed to the various detect methods to have egl/glx specific
detection for e.g. function pointer resolving.
By default egl is used if compiled with OpenGL ES, otherwise glx is used.
But in the non-gles case the windowing system can be selected through the
new environment variable KWIN_OPENGL_INTERFACE. Setting this variable to
"egl" the EglOnXBackend is used.
REVIEW: 106632
m_textureNPOT and m_limitedNPOT have not been initialized in the ctor resulting in
m_limitedNPOT being incorrectly being true sometimes for Intel drivers. For other
drivers the value had been set.
A few more getters for parsed information are added to GlPlatform.
This includes the driver information retrieved through glGetString
and information like direct rendering and loose binding.
Additionally the methods to convert version, driver and chipClass
to string are added to the public interface.
These changes allow parts outside GlPlatform to get the same debug
information about the GL system.
There seems to be a problem with nouveau GLES if you want to create an
EGLImageKHR more than once in a frame for the same pixmap. This patch
circumvents the problem in the way that it implements tfp the same way
as the mesa example in
mesa/demos/src/egl/opengles1/texture_from_pixmap.c does it. A nice
side effect of this is that it also avoids the overhead of recreating
the texture for every damaged window.
REVIEW: 103303
Fixes a major regression which allowed OpenGL
compositing to be used with software rasterizer, which
either did not work or crashed directly.
Cherry-picked from cf2f572 in branch KDE/4.7.
